Viral ‘Europe 2031’ scenario fuels AI sovereignty alarm
TL;DR:
- A speculative essay, Europe 2031, imagines the continent economically broken by 2031 for failing to build its own AI and compute capacity.
- It went viral after the US briefly blocked foreign access to Anthropic’s Fable model, lending it an air of prophecy.
- Critics note several of the eye-catching US AI deals it cites have since collapsed.
A fictional doomsday scenario is shaping real policy conversations in Brussels. Europe 2031, written by Brussels-based thinktankers, imagines a continent torn apart by 2031 — its economy gasping, populism surging and the euro wobbling — because it failed to invest in AI and the datacentres that power it while the US and China raced ahead. Published a day before Washington moved to block “foreign nationals” from Anthropic’s much-hyped Fable model, it has spread rapidly among policymakers.
A thought experiment with real-world pull
The piece belongs to a growing genre of speculative AI scenarios that have gained traction with senior officials. Its authors say it has been read by members of the European Parliament and surfaced in informal British-German official talks. They feel “vindicated” that one prediction — the US restricting global access to advanced models — briefly came true. Their prescribed remedy is blunt: Europe should build far more datacentres, faster, in deregulated “AI zones” where power and planning are streamlined.
Yet the scenario’s persuasive force sits awkwardly with its details. It name-checks a $100bn (£75bn) OpenAI-Nvidia deal that reportedly collapsed in February, a doubtful $300bn OpenAI-Oracle agreement, and Texas datacentre construction that OpenAI has since pulled out of. The authors are unfazed, presenting these as the kind of objections their fictional, doomed officials also raised. The organisation behind it, the Arq Foundation, does not disclose its funders — a detail worth weighing against its case for sweeping deregulation.
Looking forward
For the UK, the scenario lands on a genuine strategic question rather than its shakier specifics. The Anthropic access scare made plain how dependent both Britain and the EU are on US-controlled infrastructure, reinforcing Europe’s wider reckoning with that dependence and fuelling sovereign-compute efforts like OVHcloud’s frontier-AI ambitions. The harder debate — one UK experts are already having — is whether the answer is to pour money into datacentres at any cost, or to pursue sovereignty more selectively through open models and targeted capability. A viral story is a poor substitute for that calculation, even if it usefully forces the conversation.